Adhesion-dependent signaling by macrophage migration inhibitory factor (MIF).

Proper stimulation of cell cycle progression and DNA synthesis requires cooperating signals from integrin and growth factor receptors. We previously found that the proinflammatory peptide, macrophage migration inhibitory factor (MIF), functions as an autocrine mediator of growth factor-dependent ERK MAP kinase activation and cell cycle progression. We now report that MIF secretion is induced by cell adhesion to fibronectin in quiescent mouse fibroblasts. Adhesion-mediated release of MIF subsequently promotes integrin-dependent activation of MAP kinase, cyclin D1 expression, and DNA synthesis. Secretion of MIF requires protein kinase C activity, and recombinant MIF reconstitutes the activation of MAP kinases in the presence of protein kinase C inhibition. Finally, we show that cells deficient in MIF have significantly higher retinoblastoma tumor suppressor and lower E2F transcriptional activities. These results suggest that MIF is an important autocrine mediator of adhesion-dependent signaling events and may provide mechanistic insight into how MIF regulates proliferative and oncogenic processes.

Proper stimulation of cell cycle progression and DNA synthesis requires cooperating signals from integrin and growth factor receptors. We previously found that the proinflammatory peptide, macrophage migration inhibitory factor (MIF), functions as an autocrine mediator of growth factor-dependent ERK MAP kinase activation and cell cycle progression. We now report that MIF secretion is induced by cell adhesion to fibronectin in quiescent mouse fibroblasts. Adhesion-mediated release of MIF subsequently promotes integrin-dependent activation of MAP kinase, cyclin D1 expression, and DNA synthesis. Secretion of MIF requires protein kinase C activity, and recombinant MIF reconstitutes the activation of MAP kinases in the presence of protein kinase C inhibition. Finally, we show that cells deficient in MIF have significantly higher retinoblastoma tumor suppressor and lower E2F transcriptional activities. These results suggest that MIF is an important autocrine mediator of adhesion-dependent signaling events and may provide mechanistic insight into how MIF regulates proliferative and oncogenic processes.
The microenvironment of a cell plays an important role in the maintenance of normal cell morphology and gene expression. Members of the integrin family of cell surface adhesion molecules are now known to relay signals between extracellular matrix proteins in the microenvironment and intracellular signaling pathways in cells (1). The requirement for integrinmediated signaling in cell cycle progression is also well established (2). However, the exact nature of these events and how they impinge upon growth factor signaling is unresolved. A significant amount of research has focused on the role of the extracellular signal-regulated kinase (ERK) 1 MAP kinase pathway in integrin-dependent regulation of the cell cycle (2)(3)(4). The requirement for MAP kinase activation in adhesion-dependent cell cycle control is due to modulation of the expression of cyclin D and thus the activity of specific cyclin-dependent kinases (Cdks) (5,6). The regulation of G 1 phase progression relies on cyclins, Cdks, and Cdk inhibitors (7,8). Cyclin D-Cdk4/6 and cyclin E-Cdk2 activities phosphorylate the retinoblastoma (Rb) tumor suppressor, which in turn releases free E2F transcription factors, resulting in the transcription of critical S phase enzymes and regulators (9).
Despite extensive studies, the mechanism of integrin-mediated MAP kinase activation remains somewhat controversial. Although several reports have shown that activation of Ras is required for adhesion-dependent MAP kinase stimulation (10,11), others have suggested that Ras is not an essential component of this activation (12,13). Howe and Juliano (14) have demonstrated how these inconsistencies might be resolved. They report that there are two phases of ERK activation, the first being Ras-dependent and the second being Ras-independent. The first is the initial, acute phase, which requires Rasmediated Raf-1 membrane localization, whereas a sustained phase of ERK activation is independent of Ras but requires protein kinase C (PKC) activation of Raf-1 (14). Roovers et al. (15) similarly found that the sustained activation of MAP kinase is required for cyclin D1 expression, Rb phosphorylation, and G 1 phase progression in normal, untransformed cells. The authors go on to show that in order for a cell to efficiently promote sustained MAP kinase activation and cyclin D1 expression, signals from both integrins and growth factors are needed (15).
Our previous studies have established that MIF, a protein historically associated with inflammation and immune regulation, stimulates the proliferation of quiescent mouse fibroblasts (16). This response is associated with the activation of the p44/p42 ERK MAP kinases. We further demonstrated that growth factors stimulate the rapid release of preformed MIF from adherent, quiescent fibroblasts. Importantly, the sustained activation of MAP kinase in serum-stimulated fibroblasts was dependent upon MIF autocrine action (16). We now report that MIF is secreted in a PKC-dependent fashion as a consequence of cell adhesion to the extracellular matrix and plays a significant role in integrin-mediated signaling to sustained MAP kinase, cyclin D1 expression, and cell cycle progression.

EXPERIMENTAL PROCEDURES
Cell Culture-MIF Ϫ/Ϫ mice were kindly provided by Dr. John David (Harvard School of Public Health, Boston, MA) and have been described previously (19,33). MIF Ϫ/Ϫ mice and their wild-type littermates were maintained on a mixed 129Sv ϫ C57Bl/6 background (F3). MEFs were generated from embryos at day 14.5 and grown in DMEM with 10% FCS (Hyclone, Logan, UT) and penicillin/streptomycin. NIH-3T3 fibroblasts were maintained in low glucose DMEM with 10% FCS and penicillin/streptomycin. For adhesion experiments, NIH-3T3 and MEFs were G 0 synchronized for 1 and 2 days, respectively, in serum-free media. Quiescent cells were lifted from plates with trypsin-EDTA and neutralized with DMEM/2% fatty acid free BSA and 2 mg/ml soybean trypsin inhibitor. Cells then were then spun down and resuspended in DMEM with 0.2% FCS and allowed to sit for 30 min (in the presence of PKC inhibitor Ro-31-8220 (Calbiochem) where indicated). Cells were plated onto fibronectin-coated (BD Biosciences) or low cluster plates for suspension conditions (Corning Costar, Harrodsburg, KY) for the indicated times. In select experiments, recombinant MIF (rMIF) (16,17) was added to cells just prior to plating. For growth factor/adhesion experiments, cells were resuspended in DMEM with 10% FCS and plated immediately. For antibody neutralization experiments, anti-murine MIF monoclonal antibody (14.15.5, IgG 1 subclass) (18) or an isotype control monoclonal antibody was added at the indicated concentrations prior to plating of the cells. The 14.15.5 monoclonal antibody has been shown previously to neutralize both endogenously released (native) MIF and rMIF in a variety of in vitro and in vivo studies (16,19,20). DNA synthesis experiments were performed by plating quiescent, transfected MEFs in DMEM with 10% FCS or DMEM with 0.2% FCS and 20 ng/ml platelet-derived growth factor/1 M insulin (R&D Diagnostics, Minneapolis, MN) onto fibronectin-coated 96-well plates (BD Biosciences). Cells were pulsed with [ 3 H]thymidine (1 Ci/ml) (PerkinElmer Life Sciences) for 20 h. The cells then were harvested, and the incorporation of [ 3 H]thymidine into DNA was quantified by liquid scintillation counting (Packard Instrument Co.).
Immunoblotting Studies-Whole cell extracts were prepared from cells after the indicated treatments. Cells first were washed in cold phosphate-buffered saline, and then ice-cold radioimmune precipitation assay buffer (containing 1 mM NaVO4, 2 mM NaF, and a protease inhibitor mixture (Roche Molecular Biochemicals)) was added. The cells were disrupted by repeated aspiration through a 21-gauge needle. After incubation on ice for 10 min and microcentrifugation at 3000 rpm for 15 min (4°C), the supernatants were removed, the protein concentration was determined, and lysates were stored at Ϫ80°C. Equal amounts of cellular proteins were fractionated on SDS-PAGE gels (Bio-Rad) and transferred to polyvinylidene difluoride membranes (Millipore, Bedford, MA). Immunoblotting was performed with antibodies directed against phospho-ERK MAP kinase, total ERK MAP kinase (Cell Signaling Technology, Beverly, MA), p16 Ink4a , p21 Cip1 , p27 Kip1 , Cdk4 (Santa Cruz Biotechnology, Santa Cruz, CA), cyclin D1 (Upstate Biotechnology, Waltham, MA), p53 (Pharmingen), and MIF (prepared by our laboratory) (16). Densitometric analysis of Western blots was performed using the NIH Image software package.
In Vitro Kinase Assays-Whole cell extracts were prepared from 1 ϫ 10 6 cells as described above. The p44/p42 MAP kinase assay was performed according to the manufacturer's directions (Cell Signaling Technology). Briefly, equal amounts of lysate were incubated with 15 l of an immobilized anti-phospho-p44/p42 MAP kinase monoclonal antibody, and the samples were rotated overnight at 4°C. The pellet was collected by centrifugation and washed with 500 l of radioimmune precipitation assay buffer followed by three washes with 1ϫ kinase buffer. The pellet then was resuspended in 50 l of 1ϫ kinase buffer supplemented with 200 M ATP and 2 g of Elk-1 fusion protein (Cell Signaling Technology). After incubation at 37°C for 30 min, the reaction was terminated by adding 25 l of 3ϫ Laemmli sample buffer. Thirty l of each sample was electrophoresed on a 10% SDS-PAGE and transferred to polyvinylidene difluoride membrane. The blot was then probed for phospho-Elk-1 protein by utilizing an anti-phospho-Elk-1 antibody. Cdk4 kinase assay was performed similarly except that the primary antibody used was against cdk4 (Santa Cruz Biotechnology) and protein A/G beads were used to immunoprecipitate. A final concentration of 10 M cold ATP, 10 Ci of [␥ 32 P]ATP, and 2 g of the C terminus of Rb protein were added to each kinase reaction. Samples were electrophoresed, and the gel was dried and exposed to autoradiography for 12 h.
Retroviral and Plasmid Constructs-The p53 His-175 retroviral expression vectors was a gift of Dr. Oleksi Petrenko (State University of New York (SUNY) Stonybrook, NY) and is described elsewhere (16). 2 Viral supernatants were collected following transfection of viral packaging cells. 1 ϫ 10 5 fibroblasts were infected with the indicated retroviral supernatants for 24 h, washed, refed, and split 2 days later. The expression of p53 His-175 by primary fibroblasts resulted in efficient immortalization of both MIF ϩ/ϩ and MIF Ϫ/Ϫ cells as evidenced by passaging for greater than 25 times with no evidence of cell senescence. The MIF eukaryotic expression vector (pcDNA3.1 GS/MIF) has been described elsewhere (19).

RESULTS
Cell Adhesion to Extracellular Matrix Stimulates the Release of MIF from Fibroblasts-The requirement for two distinct signaling mechanisms for cell cycle progression in non-transformed cell lines is well documented (22). The first such signal requires growth factor receptor stimulation whereas the second is dependent upon cell attachment to extracellular matrix proteins via integrin receptors. It is thought that this dual signaling scheme functions to stimulate convergent signal transduction pathways resulting in maximal enzyme activation, gene expression, and cell growth (23).
We chose to investigate the role of MIF in adhesion-induced signal transduction because of our previous observation that quiescent fibroblasts held in suspension are refractory to growth factor-induced MIF release (16). To determine whether cell adhesion to the extracellular matrix induces the secretion of preformed MIF, quiescent cells, in the absence of growth factors, were either plated onto fibronectin or held in suspension (low cluster plates). Western blotting analysis of supernatants from adherent cells showed a significant amount of MIF release within 30 min of plating (Fig. 1A). In contrast, cells in suspension secreted no detectable MIF over the course of 4 h.
Autocrine Action of MIF Contributes to Integrin-induced ERK MAP Kinase Activation-The recent finding that MIF modulates MAP kinase activation in response to growth factors suggested to us that secreted MIF may also participate in signaling to MAP kinase in response to integrin ligation. The activation of MAP kinase by extracellular matrix was examined in NIH-3T3 fibroblasts as described previously (12,14,24). Quiescent cells plated onto fibronectin show a rapid activation of MAP kinase, which remains active for at least 40 min as demonstrated by immunoblotting using a phospho-specific antibody against ERK MAP kinase (Fig. 1B) and by a MAP kinase enzymatic assay (data not shown) (12,24).
It has been proposed that integrin ligation to extracellular matrix induces two distinct phases of MAP kinase activation: an acute and a sustained phase. Our previous observation demonstrated that MIF autocrine action was important for growth factor-induced sustained, but not acute, MAP kinase activation (16). To determine whether secreted MIF contributes to the sustained phase of integrin-induced ERK activation, we assessed MAP kinase activation by integrins in the presence of a well characterized, neutralizing monoclonal anti-MIF antibody. As shown in Fig. 1C, lysates from cells stimulated with fibronectin for 40 min in the presence of an IgG 1 isotype control antibody display similar levels of phosphorylated (active) MAP kinase as cells incubated without antibody. By contrast, treatment of fibronectin-plated cells with anti-MIF antibody results in a dose-dependent inhibition of integrinstimulated sustained MAP kinase activation. These data suggest that MIF plays a role in the modulation of integrininduced MAP kinase activation.
MIF Ϫ/Ϫ Fibroblasts Are Partially Resistant to Adhesiondependent Sustained MAP Kinase Activation-To better understand the role of MIF in integrin-mediated signaling to MAP kinases, embryonic fibroblasts (MEFs) were derived from MIF Ϫ/Ϫ mice and their wild-type littermates. MIF ϩ/ϩ and MIF Ϫ/Ϫ fibroblasts were examined for their relative abilities to conduct adhesion-mediated signaling to MAP kinase. As shown in Fig. 2A, MIF ϩ/ϩ fibroblasts efficiently induced MAP kinase phosphorylation at both the acute (ϳ20 min) and the sustained phase (ϳ40 min) time points (13). In contrast, adhesion-dependent MAP kinase activation in MIF-deficient cells was much less efficient than in wild-type cells. The phosphorylation status of MAP kinase in MIF Ϫ/Ϫ cells closely correlated with a decrease in ERK enzymatic activity as assessed by an in vitro kinase assay for ERK ( Fig. 2A and data not shown). The effect of MIF deficiency was less pronounced at the acute than at the sustained time point, suggesting that the contribution of MIF to integrin-mediated MAP kinase signaling is predominantly to the delayed phase of MAP kinase activation.
To exclude the possibility that the difference in integrininduced MAP kinase activation observed between MIF ϩ/ϩ and MIF Ϫ/Ϫ cells was independent of MIF loss, we reintroduced rMIF into the culture system. The addition of 50 ng/ml rMIF to MIF-deficient cells restored integrin signaling to MAP kinase (Fig. 2B) and its upstream activator, MAP kinase kinase (MEK) (Fig. 2C). We occasionally observed that the addition of rMIF to MIF ϩ/ϩ fibroblasts during integrin ligation inhibits MAP kinase activation by cell adhesion to fibronectin (Fig. 2C). We believe that this is due to a "plateau" effect: i.e. levels of integrin-induced MIF secretion are optimal for MAP kinase activation in wild-type cells, and excess rMIF increases total MIF levels into inhibitory concentrations. This effect is similar to the bell-shaped activity curves reported previously for the migration inhibitory activity of MIF (17).

Inhibition of PKC Suppresses Adhesion-dependent MIF Secretion and Sustained MAP Kinase-
The involvement of PKC in coupling integrin-mediated signaling events to MAP kinase is well documented (14,25,26). It has been proposed that PKC is required for the delayed, or sustained, phase of integrinstimulated MAP kinase (14). Moreover, several studies have suggested that PKC activity is critical for the production and secretion of MIF (27,28). To test the hypothesis that integrindependent PKC activation results in MIF secretion and subsequent autocrine activation of MAP kinase, a broad spectrum PKC inhibitor was employed. Pretreatment of NIH-3T3 fibroblasts with the PKC inhibitor, Ro-31-8220, dose-dependently inhibited adhesion-induced MIF secretion with maximal effect, displaying almost 100% inhibition (Fig. 3A). As mentioned above, PKC inhibition has been shown to disrupt the delayed phase of integrin-induced signaling to MAP kinase. If PKCdependent release of MIF is responsible for modulating the sustained activation of MAP kinase, then rMIF addition to PKC-inhibited cells should restore adhesion-stimulated MAP kinase activation. Fig. 3B illustrates the dose-dependent inhibition of MAP kinase phosphorylation by PKC inhibition, con- Lysates were analyzed for MAP kinase activation by Western blotting with total and phospho-specific antibodies against ERK MAP kinase. As shown in B, rMIF add-back restores sustained MAP kinase activation by cell adhesion. Cells were plated as in panel A for 40 min (sustained phase of MAP kinase activation), and rMIF was added at the time of plating to the indicated samples. C, as in panel B except that increasing concentrations of rMIF were added at the time of plating. In addition to total and phospho-MAP kinase evaluation, MAP kinase kinase, the upstream activator of MAP kinase, was also assessed for its activation state by a phospho-specific antibody. The data shown are representative of at least two independent experiments. sistent with previous observations (14). Supplementation of rMIF to Ro-31-8220-treated cells adhering to fibronectin fully restored the capacity of integrins to activate signaling to MAP kinase (Fig. 3B). Note that the inhibition of MAP kinase by Ro-31-8220 very closely mimics the degree of inhibition of MIF secretion (Fig. 3, A and B), further supporting the conclusion that PKC-dependent MIF secretion is partially responsible for integrin-stimulated sustained MAP kinase activation.

MIF Ϫ/Ϫ Embryonic Fibroblasts Are Deficient in Growth Factor Plus Adhesion-induced MAP Kinase Activation and DNA
Synthesis-The results described above together with our previous observation that MIF is a growth factor-induced autocrine signaling factor (16) prompted us to examine the role of MIF as a mediator of cell cycle regulators and DNA synthesis. As mentioned previously, signals derived from integrins and growth factors synergize to maximally stimulate cell cycle progression and DNA synthesis (22). It is thought that the sustained activation of MAP kinase is required for efficient G 1 /S phase progression (5,15). Because MIF is regulated by both growth factor and integrin activation and serves as an autocrine activator of MAP kinase, we hypothesized that MIF is important for growth factor/integrin-dependent signaling and cell cycle progression.
It was suggested recently that MIF is post-translationally modified and that this modification might influence bioactivity (29). To ensure bioactivity of MIF and more closely mimic a physiologic setting of MIF expression and secretion, we transiently transfected MIF Ϫ/Ϫ fibroblasts with an MIF expression plasmid to restore the phenotypic effects of MIF (19). Shortly after transfections, cells were serum-starved and then replated in the presence of 10% FCS for either 2 or 12 h. As shown in Fig. 4A, MIF status (deficiency or transfections) had no effect on growth factor/adhesion-induced MAP kinase activity at the earliest time point investigated. In contrast, growth factor/ adhesion-mediated sustained MAP kinase activation was compromised in cells lacking MIF, and reconstitution of MIF in these cells restored the defect (Fig. 4A). Note that in cells that are actively producing MIF, ectopic expression of MIF blunts the sustained MAP kinase activation, similar to what we observed after the addition of rMIF (Fig. 2C and discussed above). It should also be noted that cell supernatants from MIF-transfected MIF Ϫ/Ϫ fibroblasts stimulated with growth factors and adhesion consistently contained between 20 and 45 ng/ml of MIF, whereas MIF Ϫ/Ϫ cells transfected with the empty plasmid had MIF levels that were undetectable by immunoprecipitation/Western blotting (data not shown).
As discussed above, signals generated from cell adhesion to integrins coupled with growth factor-derived signals mediate sustained signaling to MAP kinase, cyclin D expression and, ultimately, DNA synthesis (15). Because our results suggest that MIF is important for growth factor/integrin-induced sig- naling to sustained MAP kinase, we speculated that cyclin D1 accumulation and DNA synthesis may be similarly disrupted in the absence of MIF. To determine whether MIF deficiency adversely affects DNA synthesis and cell proliferation, primary MIF ϩ/ϩ and MIF Ϫ/Ϫ fibroblasts were transfected and synchronized as in Fig. 4A. We used two different sources of growth factors to determine differences between defined and serumderived growth factors. The extent of DNA synthesis (assessed by [ 3 H]thymidine incorporation) is shown in Fig. 4B. Growth factor/adhesion-stimulated DNA synthesis of MIF Ϫ/Ϫ cells or MIF Ϫ/Ϫ cells transfected with vector alone was between 40 and 50% less than that of the MIF ϩ/ϩ cells. Importantly, reconstitution of MIF by transient transfection fully reversed the defect in DNA synthesis associated with MIF deficiency. As the defect in growth factor/adhesion-stimulated sustained MAP kinase activation is also reversed by MIF reconstitution and sustained MAP kinase activation has been shown to be important for S phase progression, these results support the conclusion that MIF contributes to MAP kinase signaling and S phase progression.

MIF Participates in Growth Factor Plus Adhesion-induced Cyclin D1 Accumulation, Rb Inactivation, and E2F-dependent
Transcription-To further delineate the requirements for MIF in cell cycle-associated events, we next sought to determine whether MIF deficiency influenced downstream targets of MAP kinase such as cyclin D1 expression and Rb/E2F activities. We recently found that immortalization of MIF ϩ/ϩ and MIF Ϫ/Ϫ primary fibroblasts with a dominant-negative mutant of p53 significantly enhances the growth differences between MIFdeficient and MIF-containing cells. 2 MIF ϩ/ϩ and MIF Ϫ/Ϫ primary fibroblasts were infected with a replication-defective virus encoding a well characterized mutant allele of p53 (p53 His-175 ) (30,31). As shown in Fig. 5A, introduction of dominant-negative p53 had no effect on the relative levels of the cyclin-dependent kinase inhibitors p21 Cip1 , p27 Kip1 (Fig. 5A), and p16 Ink4a (not shown) in normally cycling cells regardless of MIF status. In contrast, p53 His-175 immortalized MIF Ϫ/Ϫ fibroblasts displayed an impaired ability to accumulate cyclin D1 in response to growth factors and anchorage-dependent signals (Fig. 5B, upper panel). Moreover, the inability to efficiently elevate cyclin D1 levels in MIF Ϫ/Ϫ cells resulted in a proportionately reduced Cdk4 kinase activity that was not associated with alteration of Cdk4 protein levels (Fig. 5B, middle and lower panel).
Rb-mediated transcriptional repression results from decreased G 1 Cdk enzyme activities normally responsible for phosphorylating and inactivating Rb (9). Rb-dependent transcriptional repression was compared in primary and p53 His-175immortalized MIF ϩ/ϩ and MIF Ϫ/Ϫ fibroblasts by an Rb-sensitive luciferase promoter construct. Cycling, asynchronous cell populations displayed significant differences in Rb-transcriptional repression (where the degree of repression is inversely proportional to promoter activity) (Fig. 5C). MIF-deficient primary or immortalized cells consistently had higher Rb repressor activity with the p53 His-175 -immortalized cells, showing the largest differences between MIF ϩ/ϩ and MIF Ϫ/Ϫ cells. Conversely, levels of c-myc or p53-dependent transcription were unaffected by MIF status (Fig. 5C). As would be expected in cells with higher Rb tumor suppressor activity, E2F-dependent transcription was decreased in MIF-deficient cells, and the degree of reduction correlates with increased Rb-dependent repression. DISCUSSION Our prior studies revealed that MIF secretion and autocrine action contribute to growth factor-dependent sustained MAP kinase activation and DNA synthesis (16). We now show that integrin ligation to extracellular matrix stimulates MIF secretion, MIF-dependent sustained MAP kinase activation, and cell proliferation. Furthermore, cells deficient in MIF are defective in growth factor plus integrin-induced cyclin D1 expression, Rb inactivation, and E2F-dependent transcription.
Sustained signaling to MAP kinase contributes to cyclin D1 expression, Rb inactivation, and progression through the cell cycle (22). Although we demonstrate that fibroblasts derived from MIF Ϫ/Ϫ mice are partially refractory to growth factor/ adhesion-stimulated MAP kinase activation and cyclin D1 expression, it should be noted that MIF only partially regulates this phenomenon. For instance, whereas deletion of the Mif gene in mice has been achieved with two different targeting constructs, none of the resulting mice appear to suffer from developmental or growth-related abnormalities (33,34). We propose that the correct dosage of MIF in normal cells may be FIG. 5. MIF deficiency results in impaired cyclin D1 expression, Cdk4 activity and Rb/E2F function. As shown in A, primary MIF ϩ/ϩ and MIF Ϫ/Ϫ embryonic fibroblasts were immortalized by infection with a replication-defective retrovirus encoding a dominant interfering mutant of p53 (p53 His-175 ) as described under "Experimental Procedures." Three days after infection, cells were analyzed for relative levels of p53, p21 Cip1 , p27 Kip1 , and Cdk4 by immunoblotting. wt, wildtype; ko, knockout. As shown in B, quiescent, p53 His-175 immortalized MIF ϩ/ϩ and MIF Ϫ/Ϫ MEFs were plated onto fibronectin-coated dishes with DMEM and 10% FCS for 6, 9, 12, and 15 h. Cell lysates were subjected to immunoblotting (IB) for cyclin D1 and Cdk4. A parallel set of samples was lysed, and Cdk4/cyclin D complexes were immunoprecipitated and incubated in a kinase reaction with glutathione S-transferase (GST)-Rb peptide as a substrate as described under "Experimental Procedures." As shown in C, primary MIF ϩ/ϩ and MIF Ϫ/Ϫ embryonic fibroblasts and p53 His-175 immortalized MIF ϩ/ϩ and MIF Ϫ/Ϫ fibroblasts were transiently co-transfected with the indicated transcription factorresponsive luciferase plasmids and the Renilla pRL-TK vector for 40 h. Results are expressed as fold increase over control after normalizing ratios of luciferase/Renilla luciferase from quadruplicate samples. All experiments were repeated at least twice with similar results. important for optimal cell growth under specific conditions and that in immortalized or transformed cells, there is a greater requirement for an MIF-dependent contribution to MAP kinase activation and cyclin D1 expression.
Despite extensive efforts, a classical, membrane-bound receptor for MIF has not yet been described. An intriguing mechanism by which MIF may carry out its cellular actions was recently proposed by Kleemann et al. (35). They describe that extracellular MIF has the unique ability to traverse the plasma membrane and interact with cytosolic c-jun-activating binding protein (Jab1). Jab1 effectors include the transcription factors AP1, HIF1␣, and SMAD4, among others (36,37). Another interesting effector reported for Jab1 is the ␤ 2 subunit of integrin LFA-1 (38). This raises the possibility that MIF binding to intracellular Jab1 could influence the dynamics or strength of integrin signaling cascades. Whether the MIF/Jab1 interaction is the sole mechanism by which MIF exerts its pluripotent activities will require more thorough investigation.
Welsh et al. (39) recently described a crucial role for Rho GTPase in modulating sustained activation of MAP kinase and the appropriately timed expression of cyclin D1 in mid-G 1 phase. This study reaffirmed the importance of MAP kinase and the timing of cyclin expression for progression through the cell cycle while highlighting a novel functional role for Rho in cell growth regulation. Although beyond the scope of this manuscript, elucidation of a role for MIF in the modulation of Rho or vice versa would yield critical insight into the biology of MIF in cell cycle regulation.
Hudson et al. (40) recently reported that MIF has the ability to negatively regulate p53-dependent processes. In a cell-based genetic screen for modulators of p53 function, MIF was identified and found to suppress p53-dependent cell senescence and apoptosis. Although our studies support this finding, they additionally support the idea that MIF function modulates Rb/ E2F-dependent processes and that MIF regulation of these two pathways may, in fact, be linked. 3 In a separate study, we find that MIF action in cellular transformation impinges upon the Rb/E2F pathway, but the net cause and effect of this remain unclear. 2 From these studies we cannot conclude that the blunted signaling to MAP kinase and cyclin D1 expression in MIF Ϫ/Ϫ cells is, in fact, causal to the transformation resistance we have observed. Although it is premature to conclude that MIF promotes cell transformation by modulating cyclin D1 expression and Rb/E2F activities, it is not unreasonable to speculate that MIF-dependent regulation of sustained MAP kinase activity may be an important contributing factor in oncogenesis.
The importance of cytokines and growth factors in maintaining tumor growth and viability is well established (41). This field of study has recently become the focus of several novel targeting strategies for anticancer chemotherapies (42). In particular, the family of epidermal growth factor ligands and receptors has gained much attention because inhibitors of the epidermal growth factor receptor tyrosine kinase and ligandneutralizing antibodies have shown therapeutic efficacy in preclinical and clinical trials (43). Interestingly, dysregulation of this ligand and/or the expression of its receptor results in persistent activation of MAP kinase, cyclin D1 expression, and tumor promotion (44). Together with prior studies showing the antitumorigenic effects of MIF inhibition (45,46), the present data suggest that MIF may represent an important new target for cancer chemotherapeutics.